INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.90

ulner

-0.79

isd

-0.78

phrine

-0.73

ouble

-0.73

anamo

-0.71

oub

-0.69

udeb

-0.69

umbledore

-0.68

plom

-0.67

POSITIVE LOGITS

 Punk

0.67

itud

0.66

 Bees

0.65

 Raider

0.63

 Animals

0.62

Tools

0.59

 Sergeant

0.58

Drop

0.57

igator

0.57

bags

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.